​

VegaGraph type catalog

The type catalog is the schema system behind VegaGraph. Built-in resources are loaded from checked-in catalog definitions; authorized users can work with tenant-visible types through the public type-catalog APIs.

​

Entity types

Entity types use a thin entity_schema for identity and display fields. Rich metadata belongs in reusable aspects. Built-ins cover data, software, infrastructure, business, governance, operations, and knowledge resource families.

​

Aspect types

Aspect types provide strict JSON Schema for fact payloads. They can also identify searchable fields and UI presentation metadata. Examples include description, ownership, lifecycle, classification, compliance, observability, runtime, cost, and data quality. An entity accepts only aspects bound by its entity type, and every aspect payload is validated before persistence.

​

Relationship types

Relationship types define source and target constraints, endpoint granularity, metadata schema, cardinality, acyclic behavior, and transitivity. Choose these rules deliberately: they affect valid writes, traversal, lineage, and impact.

​

Naming and evolution

Use stable snake_case names. Prefer adding a reusable aspect over widening every entity identity schema. Treat schema changes as compatibility decisions because existing entity, aspect, and edge data must continue to validate and remain understandable.

​

Permissions

Entity types, aspect types, and relationship types have collection and instance permission surfaces. A reader may use a built-in type without being authorized to define or change types. Secret values are references only. A type may model a reference or classification, but its schema must never invite storage of the credential itself.